> -----Original Message-----
> From: David [mailto:david@drccc.org.au]
> Sent: Thursday, January 13, 2005 6:12 PM
> To: squid-users@squid-cache.org
> Subject: [squid-users] digest
>
>
> Hi All,
>
> I've just set up Squid on a Debian Sarge box and tried to configure it
> so that it will proxy/cache for my local network which is connected via
> ISDN to my ISP. When I run Squid, and make the first request for a
> webpage, it appears to go off and try to get the file
> /squid-internal-periodic/store_digest from the nominated parent (ISP
> runs Squid 2.5.). This file is some 6MB is size and uses most of my
> link's resources to download. I am not sure what happens next but I see
> a continual transfer of data at about 6MB per hour (in darkstat) making
> me think that the digest file is constantly being downloaded.
>
> I don't really need the digest anyway as I only have one access point to
> the wider world and no other sibling caches, just the ISP parent. I've
> tried various squid.conf configurations and nothing works right. I
> either get it working with the parent and downloading my wanted webpages
> BUT along with the %&# digest OR my proxy tries to access the wanted
> webpages directly and the ISP firewall stops that.
>
> Any way I can get squid to simply act as a local caching proxy, getting
> non-locally stored stuff from a single parent proxy without dwonloading
> the digest almost contunuously? Has anyone else seen this digest info
> continuously flowing? I am assuming that the problem is mine as I have
> another box running Squid 2.2.STABLE6 and it works fine.
>
> Thanks for any help.
> David...
On your cache_peer line, specify no-digest as one of the options. For
example:
cache_peer proxy1.example.com parent 3128 3130 no-digest no-query
Since your setup doesn't allow you to grab pages directly, there is no point
in querying your parent cache to see if it has the pages (either you have
them locally, or you are going to get it from the parent cache), hence the
no-query option. Also be sure to use "never_direct allow all" to prevent
your cache from trying to fetch web pages directly.
Chris
Received on Fri Jan 14 2005 - 11:54:12 MST
This archive was generated by hypermail pre-2.1.9 : Mon Mar 07 2005 - 12:59:35 MST